Thesis/Capstone
Publication Date
Authored by
Sandra Rhee, Bernardo Garza
Abstract

Our sponsor seeks to transform negotiation into a data-driven process while preserving institutional knowledge of category sourcing strategies. To support this goal, we developed the Negotiation Buddy, an intelligence tool designed to help sourcing professionals identify key negotiation levers and generate effective counteroffers. The tool integrates sourcing practices such as Kraljic’s Matrix and Porter’s Five Forces into a structured, analytics-driven workflow. Through an intuitive dashboard, users receive tailored recommendations based on internal value, risk assessments, and external market dynamics. Historical rate card data is processed using machine learning to uncover patterns in supplier behavior. We applied K-Means clustering with category-specific features (e.g., supplier, role, and experience level) and historical rates. The analysis revealed an optimal clustering of three groups, each demonstrating distinct pricing and negotiation behavior. To predict supplier pricing, we trained an XGBoost regression model that achieved strong accuracy (MAE = $6.80, MAPE = 9.38%). SHAP analysis clarified the contribution of controllable (e.g., contract terms, timing) and uncontrollable factors (e.g., inflation, labor cost shifts) on pricing. The pilot was conducted within the sponsor’s procurement function for application development services in China, using rate cards from 11 suppliers across 45 roles and three experience levels from 2020, 2022, and 2024. The tool includes a predicted rate card to serve as a benchmark and an analytics module that surfaces key supplier insights and category trends. By capturing and learning from negotiation outcomes, the Negotiation Buddy facilitates institutional learning and continuous improvement, helping sourcing teams to negotiate more strategically, consistently, and at scale.

Attachment(s)